AI Models

Training compute growth

Estimated training compute of frontier models by year (Epoch AI).

← Back to The Frontier AI Model Race (2026)

About this data

Estimated training compute for frontier models has grown several orders of magnitude, on the order of 4–5× per year per Epoch AI. The log axis makes the otherwise-exponential trend readable and underlines why power and chips have become the binding constraints.

Training compute growth

Estimated training compute per frontier release (log scale, FLOPs).

View data & sources →

Data table

Training compute growth — training_compute data table (The Frontier AI Model Race (2026))
year series source_ref value_basis flops_estimate
2020 training_compute epoch-ai GPT-3 ~3.1e23 FLOP (Epoch AI; corroborated by Stanford HAI AI Index) 3.1e+23
2022 training_compute epoch-ai PaLM-class frontier ~2.5e24 FLOP (Epoch AI) 2.5e+24
2023 training_compute epoch-ai GPT-4 ~2e25 FLOP (Epoch AI) 2e+25
2024 training_compute epoch-ai Gemini Ultra ~5e25 FLOP (Epoch AI) 5e+25
2025 training_compute epoch-ai Grok-4 ~5e26 FLOP, largest known run (Epoch AI) 5e+26

Methodology & sources

Last updated: Jul 17, 2026

Methodology

Source-backed values are seeded for four of the five charts: the release cadence by lab (2024 → 2026, from each lab’s release notes), the capability-vs-cost scatter (Artificial Analysis Intelligence Index vs output $/M tokens, June 2026 snapshot), training-compute growth by year (Epoch AI, corroborated by Stanford HAI), and a safety-gated capabilities matrix. Every numeric point carries a sources[].ref and a value_basis. Sources are each lab’s own model cards/release notes, Artificial Analysis, Epoch AI, and Stanford HAI — cross-checked against public release timelines. EDITORIAL ENCODING: the safety-gated matrix scores each model × domain as 3 = allowed / 2 = gated / 1 = blocked. This is an interpretation of each model’s published safety policy, not a measured benchmark. PLACEHOLDER: the per-lab benchmark-trajectory chart is left unseeded — a consistent historical per-quarter, per-lab benchmark series was not sourceable without mixing incompatible benchmarks. Re-verified 2026-06-15.

Sources

Comparisons are informative, not definitive. See each source for definitions and limits.

Continue researching

How relevant was this information?